ok, this may sound stupid, but I was doing a remote install of gentoo, with no physical access to the machine at all. there was a debian installed with a swap partition of 2G which I used a root partition for a gentoo chroot. anyway, when I finally thought I was ready, I rebooted the machine to the new gentoo kernel. waiting for a ping reply I suddenly realised that, even if the machine came up again, I wouldn't be able to ssh to it, because I didn't do # rc-update add sshd default oh well, I just payed for an intervention there, to reboot the machine back into the debian (whose kernel and root I luckily set as a second option in grubs menu.lst) and fixed the problem. so here's my remark: maybe we should put an explicit note in the installation document to avoid these kinds of mistakes in the future. (maybe also add that when PermitRootLogin is set to False that at least one user is to be added to the wheel group too).
(In reply to comment #0) We already tell you to add a user to the wheel group. And we already talk about starting up sshd -- we expect that if you're going to be doing something advanced like a remote install, that you keep track of what programs you'll need to emerge. Unfortunately, we can't include warnings for every possible bit of software a user might include, because if we wrote things like "oh, if you emerged *foo*, don't forget to configure it like so..." the handbook would be another hundred pages long.
yeah, well... ok then. :-) it's not all that important, I was just so frustrated by my mistake that I didn't want others to overcome the same. anyway, thanks for the answer.
It's beyond the scope of the regular handbook. I'd expect users doing such an install to be more familiar with Gentoo and they would be more likely to use the quick install guide. http://www.gentoo.org/doc/en/gentoo-x86-quickinstall.xml